howscrewedami
Joined 9 years ago
Joined 9 years ago
> Electron no longer allows Node.js to be called from renderer processes, all communication with Node.js is done via IPC.
I don't think calling Node.js from the renderer was a huge selling point for Electron - as it is not recommended due to security concerns. However, I still believe you can achieve this if you want - by using contextIsolation: false and nodeIntegration: true. But, even if this is not possible, it looks to me that using IPC with contextBridge still allows for pretty flexible Node.js use.
> Why does it have to be tied to V8/Node.js?
Some people prefer to have a Node.js backend (backend here being Electron's main Node.js process). There are multiple advantages to this:
- If using typescript: shared types between the frontend and backend
- Shared utilities, functions and constants
- If utilities and functions are shared, that means we can write some tests once, and the tests will work for both the backend and frontend
- Less context switching
These are just a few, I'm sure there are more.
> In this case, why do we still need Electron?
Besides the reasons I listed above, the alternative you listed (Tauri) seems to be relatively newer, have a smaller community, and with less online resources available (tutorials, stack overflow questions, etc).
"To allow GPTBot to access your only parts of your site"
https://openai.com/policies/terms-of-use
The only thing I found was:
You may not (...) (vii) buy, sell, or transfer API keys without our prior consent.
Apologies for sounding like an absolute nerd, but you can go with options 1 and 3 - as long as you ask OpenAI first and get their consent. Also the meaning of "transfer" is a bit murky in my opinion. If I take my API key and write it down in a napkin, technically I've just transferred my API key to a napkin - sorry Mr. Altman!
I strongly disagree. It actually does tell you quite a lot. For example, if there aren't any signals, you're very likely not seeing anything.
> The only way to understand that you’re seeing an apple from this signal is to train a model on your responses
It seems to me this contradicts your first argument. If such models exist (they very likely do, I'm not familiar with this area of research), they can tell us a lot about a person's experience.
> see “pholosophical zombie” concept.
Searched for “pholosophical zombie” but couldn't find anything, sorry. Just kidding :)
> There's no way to know if/what an AI experiences
Getting the state of neural net, at a given point in time, is easy. There are many ways to see exactly which neurons activate, why they activate, how much they activate, etc. For smaller neural nets, this is actually easy to do - here's a blog post about it:
https://distill.pub/2020/circuits/visualizing-weights/
As neural nets get larger and larger, interpretability gets harder and harder. However, I wouldn't say it's impossible.
That is quite an absolute statement. Some people just have a personal preference for having git trees arranged in a certain way. If squashing can help with that, then it serves a purpose.
> let’s assume that the utility gain for the gifted son from living in the suburbs would be larger than the utility gain for the disabled son from living in the city. A pure utilitarian, then, must choose the suburbs.
Everything's okay so far. But then he says this:
> Nagel’s view is this: if you say that you would live in the city for the sake of your disabled son, despite it being the case that moving to the city creates more utility in total, you are not a utilitarian
Didn't the author just say moving to the suburbs creates more utility in total?? And now he's saying moving to the city is what creates more utility?
You can get a lot of useful information by only saving stuff on the server. e.g.: number of visits per day, user behavior on certain pages, etc etc...